Page:EB1922 - Volume 31.djvu/220

190 is required for this purpose. In McClelland's experiments the ion zation might have been into positive and negative ions rather tha into positive ions and electrons; before the negative ions could b efficient for ionization by collision they would have to undergi further dissociation into electrons and uncharged molecules. Curve similar to that in fig. 6 have also been obtained by O. W. Richard son. Pawlow (Proc. Roy. Soc. A. 90, p. 398) and also Franck and h. v. Bahr ( Verh. d. Deutsch. Phys. Ges. xvi, p. 57, 1914) came to th< conclusion from their experiments, that ionization was producec by positive ions even when their energy did not exceed a few volts indeed they could not get any evidence of a minimum to the ionizing voltage. Horton and Davies (Proc. Roy. Soc. 95, p. 333) could no detect any ionization in a gas by positive helium ions when the energy was due to 200 volts. They ascribe the ionization observed by Pawlow and Bahr and Franck to photo-electric effects; they con sider, however, that positive helium ions can liberate electrons frorr a metal against which they strike if their energy exceeds 20 volts Baerwald considers that it requires an energy measured by 900 vojts before positive ions can liberate electrons from metals.

There are thus at least four methods by which the supply of elec- trons near the cathode necessary to maintain the discharge can be obtained. The gas near the cathode may be ionized by positive ions or by radiation, or the cathode itself may emit electrons under the impact of positive ions or by the incidence of radiation.

When the gas is at a low pressure the appearance of the discharge has well-marked characteristics which may throw light on the method by which the electrons are produced and the place from which they start. The discharge near the cathode is representec in fig. 7; near the cathode we have a velvety glow, then a space comparatively dark called the cathode dark space ; this joins on to

a brightly luminous region called the negative glow; passing through this region, and making themselves evident by the luminosity they excite when they strike against the glass wall of the vessel in which the gas is contained, are the cathode rays. These have been shown to be electrons moving with high velocity. These electrons have been liberated by the action of the electric field and have acquired their velocity under the action of that field. The velocity of the cathode rays has been measured, and it has been found that practi- cally all of them have the same velocity. This shows that they must nave all fallen through the same potential. They would do this if they all started from the cathode itself, but if they had originated by the ionization of the gas in the dark space in front of the cathode some would have started from one place and some from another' and they would have acquired different velocities. This is strong evidence in favour of the cathode itself being the primary source of the electrons which maintain the discharge. When a supply of electrons is produced by processes taking place at the cathode ionization by collisions of electrons with the molecules of the gas is sufficient to maintain the discharge through the interval between the negative glow and the anode. This interval, as will be seen from fig. 7, is made up of a short part next the negative glow in which there is comparatively little light, called the Faraday dark space, and then a long uniform portion reaching right up to the anode. Unless the pressure is very low or the spark very short this position, which is called the positive column, forms by far the larger part of the discharge. The discharge here will be maintained if the rate at which electrons are produced by collision is equal to the num- ber lost by recombination. When this is the case, equation (19)

gives o=7/>, or, since a is of the form pf-

/

= 7,

thus XeX=cQ, where c is a quantity which does not depend upon the pressure or strength of the field; as X is inversely proportioned to the pressure, this equation is equivalent to X=c,p, when d is a quantity which will depend on the nature of the gas and possibly on the intensity of the current. If / is the length of the positive column the difference in potential between the anode and the end ot the positive column next the cathode is IX, i e Idp

Between the cathode itself and the negative glow there is a fall of potential, called the cathode potential fall, which, when the cur- rent carried by the discharge is not large, is independent of the current and the pressure of the gas; it depends upon the nature of the gas and the material of which the electrodes are made. If V is

le cathode fall then (neglecting the change in potential in the negative glow and the Faraday dark space, which has been found

by experiments to be very small) V, the potential difference between the anode and cathode will be given by the equation

V = V,+Cilp (21).

It is assumed that the length of the spark is greater than that of the dark space D : at pressures comparable with that of the atmos- phere, D IF a very small fraction of a millimetre, but at the low pressures which can easily be obtained in highly exhausted vessels D may be several centimetres. It is to be noticed that V is a linear function of Ip, and Ip is proportional to the mass of gas between the electrodes; hence as long as the mass of gas between the electrodes remains unaltered the potential difference required to maintain the spark will be constant. This law, which was discovered by Paschen in 1889 as the result of a long series of experiments, is known as Paschen s law." It has been found to be in agreement with the very numerous investigations which have been made on the poten- tial difference required to produce a discharge in an approximately uniform electric field such as that which exists between two slightly curved electrodes.

The relation (21) does not give any indication of the relation between the potential difference and the spark length when the latter is exceedingly small. When the spark length falls below a "i 1 - u va . e "hich is inversely proportional to the pressure, and which in air at atmospheric pressure is about -01 mm., the spark potential increases rapidly as the spark length diminishes; this was first observed by Peace. A simple way of demonstrating it is to use slightly curved electrodes and to observe the path of the spark as these are brought closer together. Until the electrodes get very close together the spark passes along the shortest line between them, but as they approach each other a stage is reached where the spark no longer passes along the shortest line but goes to one side, taking a longer path, showing that it is easier to produce a long spark than a short one. The relation between the potential difference and the spark length for several gases has been deter- mined by Carr, who finds that Paschen's law that the potential difference depends only on pd is also true for very short sparks; 1 aschen s own experiments were made with sparks considerably longer than the critical value. Fig. 8 represents Carr's results for

2000 1800 1600 MOO 1200 1000 800 600 4OO

200

12*4567

Product of Pressure and distance betwetn Electrodes o.

FIG. 8

he relation between V and pi. The results of Carr and Strutt's experiment for the minimum spark potential, and the value of pi, at which it occurs, are given in the following table:

Minimum Spark

Potential in Gas. volts.

A'. r 341 S

Nitrogen 251 S

Oxygen 455 C

lydrogen j 3 ^^ 9 f

arbonic acid 419 C

Sulphur dioxide 457 C

Citrous oxide 418 C

Sulphuretted hydrogen 414 C

Acetylene. . . 468 C

Helium 2 6l S

he curves are very flat in the neighbourhood of the minima""so hat the critical values of pi may be subject to considerable er- ors. Strutt found that even very small traces of impurity pro- luced very large increases in the values of the minimum spark lotential in nitrogen and helium ; these are gases where, as we ave seen, such traces produce large diminutions in the mobility f the negative ion. The existence of a minimum for the spark jotential and a critical spark length follow from the view that he spark is maintained by the emission of electrons from the athode owing to its bombardment by positive ions. For if i be the number of cathode rays emitted from unit area of the athode per second, at a distance x from the cathode ioe elec- rons will stream through unit area per second and will produce

pi.

5-7 6-7

14-4

5-1 3'3 5 6

35